Low-cost revamp strategies improve FCCU performance

Sept. 24, 2007
Hindustan Petroleum Corp. Ltd. (HPCL) used low-cost modifications to the FCC unit in its Mumbai, India, refinery to improve performance.

Hindustan Petroleum Corp. Ltd. (HPCL) used low-cost modifications to the FCC unit in its Mumbai, India, refinery to improve performance. The low-cost revamp, which cost about $2 million, generated $0.2/bbl in savings with a payback of 5 months.

The FCC is the primary source of gasoline and LPG production in many Indian refineries. Increasing unit capacity and improving yield and product quality have a dramatic effect on refinery economics and the FCC’s ability to help meet new product specifications.

The FCC process has complex interactions between feed, catalyst, and hardware. Identifying the hardware constraints in FCCUs and unlocking their potential using simple and low-cost modifications enhance cracker capacity and yields.

The FCC unit at HPCL’s Mumbai refinery has a 20,000 b/sd, side-by-side reactor-regenerator. The key constraints were catalyst circulation, reactor stripper head, orifice chamber, wet gas compressor, and downstream facilities.

The 2005 revamp involved stripper internal modifications with an elevated bed, catalyst transfer-line aeration changes, a redesign of the orifice chamber, modifications in the regenerator air grid, changes in fractionator internals, and the installation of a high-efficiency rotor bundle. These modifications, which required little investment, have unlocked the FCC’s potential by improving its throughput and yields.

Background

HPCL’s original FCCU with bed cracking was commissioned in 1955. The unit was revamped in 1999 to modernize the facility and to increase the feed rate to 20,000 b/sd from 10,000 b/sd.

The original FCC licensor conducted the basic engineering design for the revamp (reactor, regenerator, transfer lines, etc). The revamp design included some of the licensor’s new technologies, such as new high-pressure-drop feed injectors, new catalyst transfer lines, extended feed riser terminating at cyclone inlet elevation, etc.

The revamp design was, however, constrained because the existing unit was to be modified while maintaining the basic vessel layout, structure, and foundation. The refiner designed the downstream sections including the waste-heat boiler, stack valves, orifice chamber, cat fractionator, and gas concentration unit (GCU) section.

The modernized FCC had advantages compared with the earlier bed-cracking unit:

  • High temperature regeneration (690-720° C. vs. 600° C.) led to less carbon on the regenerated catalyst.
  • Riser and transfer-line cracking vs. dense bed cracking improved product selectivities.
  • A cold-wall reactor design improved mechanical reliability.
  • Advanced high-pressure drop feed injectors vs. open pipe injectors gave a better yield pattern.

The changeover from the existing bed cracking to a new FCC design improved process unit yields substantially. Because the actual feed quality was quite different from the design feed, however, the resulting yields were different due to changed operating conditions. Subsequently, HPCL and the FCC licensor worked together to analyze and resolve the limitations, which helped to improve the unit operation substantially.

Click here to enlarge image

Table 1 shows unit performance in 2004 vs. design conditions. There was still a gap between the design and actual yields and we felt we could improve unit performance. Unit constraints in 2004 were:

  • Suboptimal catalyst circulation resulting in low catalyst-oil ratios and high heater outlet temperatures that led to undesirable production of gas and coke.
  • Limitations in regenerator air plate grid and the flue gas stack valves and orifice chamber.
  • Limitation in the desuperheating section in the main fractionator.
  • Wet gas compressor capacity.
  • GCU capacity and product quality.

Problem identification

For a pressure-balanced unit, such as the one at Mumbai, catalyst circulation rate is controlled and adjusted with changes to the spent-catalyst circuit. The primary control variables are unit differential pressure, stripper level, and control air rate (to a lesser extent).

Click here to enlarge image

The regenerated-catalyst circuit then responds to these changes by adjusting the overflow-standpipe catalyst level to maintain the catalyst circulation-rate constant between the two vessels. Table 2 shows the three spent catalyst-circuit control parameters that increase catalyst circulation.

Once the catalyst level approaches the top of the overflow standpipe in the regenerator, the FCC becomes nonresponsive to additional changes in the spent-catalyst circuit and no further increases in catalyst circulation rate are possible. Typically, the stripper level will decrease because the amount of catalyst returned to the reactor is less than that transferred to the regenerator. Optimization of both transfer-line circuits is therefore necessary to achieve maximum catalyst circulation.

Click here to enlarge image

Table 3 shows the results of a detailed pressure survey that compares the design and actual pressure profile. The survey showed these limitations:

  • Pressure buildup in the reactor stripper was lower than design.
  • Pressure drop in spent-catalyst transfer line was higher than design.
  • Pressure buildup in regenerator standpipe was high. Due to low catalyst density in standpipe (20 lb/cu ft), however, the standpipe was essentially full with catalyst to the top of the overflow standpipe.
  • Pressure drop in the regenerated-catalyst transfer line was higher than design.

In 2005, to improve unit performance in the constrained areas, we modified the reactor stripper, orifice chamber, air grid, wet gas compressor, main fractionator, GCU, and catalyst formulations.

Stripper modifications

To increase catalyst circulation, we raised the pressure in the spent-catalyst circuit by raising the catalyst level in the stripper. With a reactor stripper level of 70%, pressure at the bottom of spent-catalyst circuit was about 4.2 psi. This was just enough to overcome the pressure drop in the catalyst transfer lines and spent-catalyst riser for a circulation rate of 7.0 tonnes/min.

Increasing circulation to design level of 9.3 tonnes/min required a pressure of 6.0 psi in the reactor. To achieve this, we proposed increasing the catalyst height in stripper by 3 m.

To avoid channeling of catalyst and stripping steam and defluidization of catalyst at the increased stripper height, we added a new stripper shed cone close to the new catalyst level. This cone was attached to the riser and had multiple holes at the periphery to allow for good steam and catalyst contact.

Fig. 1 depicts the new stripper shed attached to the reactor riser.

Click here to enlarge image

The increase in catalyst level caused the reactor diplegs to be submerged in the catalyst bed. We evaluated the effect of increased catalyst level on cyclone performance and determined that there was sufficient dipleg outage to satisfy the cyclone pressure balance. We therefore made no changes to the cyclones.

Standpipe catalyst density

During earlier pressure surveys of the regenerator standpipe, the catalyst density was very low at 20-22 lb/cu ft. This was comparable to a regenerator bed density of 18-20 lb/cu ft.

Aeration along the regen standpipe is one of the hardest things to control in an FCCU because overaeration and underaeration can lead to low-pressure buildup in the standpipe and, ultimately, limit catalyst circulation. If the standpipe is underaerated, the pressure buildup will be low and frictional losses will be high. Conversely, if the pipe is overaerated, bubbles will form that will restrict catalyst flow.

The volume of aeration required along the length of pipe depends on the catalyst properties, standpipe conditions, distance between the aeration points, and catalyst circulation rate. In practice, however, the field transfer line’s pressure surveys indicated an inconsistent pressure profile. We rectified this in gradual steps by changing the aeration pattern. This improved the catalyst circulation from the regenerator to the reactor.

Changing the aeration pattern and increasing the pressure in the stripper improved overall catalyst circulation rate and stability.

Orifice chamber

One of the tools for an operator to control catalyst circulation in a pressure-balanced unit is the pressure differential between the reactor and regenerator. Reactor pressure is usually fixed based on fractionator pressure and the regenerator pressure is varied to increase or decrease catalyst circulation.

Two sets of double-disc stack valves, also called pressure differential control valves (PDRCVs), control the regenerator pressure.

Flue gases exiting the regenerator at 730-740° C. cool to 220° C. by generating medium-pressure steam in a waste-heat boiler located in the flue gas line. Cooled flue gas then flows through the two PDRCVs and then two fixed-restriction orifice plates (RO1 and RO2) before discharging through the stack to the atmosphere.

Fig. 2 shows the flow of regenerator flue gas.

Click here to enlarge image

Well-operating PDRCVs are necessary for having fine pressure control between the reactor and regenerator and, therefore, steady catalyst circulation and reactor temperature. Since initial unit commissioning, the PDRCVs only opened 1-2%. To keep them in a controllable range, the airflow rate to regenerator was kept higher than was required; this led to high catalyst losses.

Small PDRCV openings indicated that the fixed orifice plates had a small pressure drop and almost the entire system’s pressure drop (90%) occurred across the PDRCVs. This was affecting unit stability because the catalyst circulation rate was erratic and led to swings of 8-10° C. in the reactor temperature.

We therefore redesigned the orifice plate’s assembly to take a 75:25 pressure drop across the orifice plates vs. PDRCVs. The existing two-plate orifice assembly was replaced with a three-plate assembly. Each orifice plate was designed with a greater pressure drop vs. the earlier operation.

We increased the orifice plate’s thickness to be consistent with a maximum design pressure drop of 12 psi. A total of 30 holes were provided in each plate with 22, 26, and 29 open holes, respectively.

Click here to enlarge image

Table 4 shows the pressure profile with the two- and three-orifice-plate systems.

The revised pressure profile increased the PDRCV opening to 15-20%, which led to these advantages:

  • Better unit stability. The catalyst’s circulation rate is steadier and reactor temperature swings reduced to 1-3° C.
  • Reduced airflow to regenerator air grid. When the pressure drop across the orifice plates increased and the PDRCVs opened, airflow was lowered closer to the theoretical requirement. This reduced catalyst losses substantially, to 1.3 tonnes/day from 2.5 tonnes/day.

Afterburning, air grid

The new FCC, commissioned in 1999, was originally designed to operate in partial-burn combustion mode with a CO concentration of 6 mole %. Due to a big difference between the design and actual feed-quality basis, however, the FCC currently operates in full-combustion regeneration mode (excess flue gas oxygen).

Generally, for a full-burn regenerator, operating the dense bed at more than 700° C. and maintaining the flue-gas excess oxygen content at 1-2% can minimize afterburning. This ensures that sufficient oxygen is available to complete combustion in the dense bed.

At oxygen levels of less than 1%, the unit will potentially operate in the full burn-partial burn transition regime, resulting in afterburning. Refiners must avoid this transition regime by closely monitoring the flue-gas analysis.

Although the use of CO promoter can be effective in controlling afterburning, we still experienced severe afterburning, which indicated that it resulted from CO and oxygen breakthroughs from the dense bed.

When we mapped the regenerator’s temperature profile, we observed that afterburning was more severe in the spent-catalyst riser section. The secondary cyclones’ outlet temperatures on the spent-catalyst riser side were 50° C. higher than the cyclone outlet temperatures relative to the overflow well side.

The temperature profile indicated CO was breaking through from the spent-catalyst riser section and oxygen was breaking through from the overflow well section. Excess CO was combining with oxygen in the plenum section and across the cyclones, leading to severe afterburning.

We revised the regenerator grid hole pattern to have more grid holes in the spent-catalyst riser side to provide more oxygen. The final pattern increased the ratio of holes in spent-catalyst riser/overflow side to 60/40 vs. a typical design ratio of 55/45. This change reduced afterburning significantly.

The cyclone outlet temperatures in the regenerator plenum now closely match. Regenerator dense-bed temperatures near the spent-catalyst riser increased 10° C., which shows that there is better combustion on the spent-catalyst riser side.

We offset the main air blower capacity limitation expected with increased airflow to the regenerator air grid by increasing the number of air grid holes to 790 from 730. The pressure drop across the grid dropped to 2.2 psi from 2.7 psi; this lower backpressure to the air blower helped increase air to the regenerator by 2,000 scfm after the 2005 revamp.

Wet-gas compressor

The existing wet-gas compressor (WGC) was earlier revamped to produce a high polytropic head with a capacity of 21.5 tonnes/hr. The compressor spillback valve controls the fractionator pressure.

A study of the reactor-regenerator section indicated that with an increase in catalyst circulation to 9.0 tonnes/min, and with a new catalyst formulation, wet-gas generation can increase to 30 tonnes/hr from 21.5 tonnes/hr. A check of the WGC at 30 tonnes/hr indicated that both the compressor and the motor were limiting.

We took this approach to debottleneck the WGC:

  • Reduce the number of impellers to lower the compressor’s polytropic head. The current rotor bundle had six impellers and was developing 41,000 ft/lbf of head, which could be reduced to capture the capacity enhancement.
  • Increase the compressor’s suction pressure. In the existing system, there was about 14 psi of pressure drop across the suction control valve to meet the suction pressure requirement of existing impeller. Increasing suction pressure raises the mass flow through the compressor, therefore increasing capacity and reducing the polytropic head.
  • Replace the existing low-efficiency rotor bundle with a high-efficiency rotor bundle consisting of diaphragm, guide vanes, valves, and labyrinths. This can reduce internal leakages and compressor efficiency could increase to 79.5% from 66.5%.

New hardware developments enabled us to revamp the WGC with a five-stage rotor bundle. It avoided the need to procure a completely new compressor, which would have required substantial changes in the unit’s structure and foundation, changes with associated inlet and outlet piping and accessories, and higher costs.

A lower polytropic head and greater compressor efficiency enabled the existing motor to handle more wet gases without any modifications.

Click here to enlarge image

Table 5 compares conditions before and after the revamp.

Main fractionator

Modifications in the reactor-regenerator and WGC created the potential to increase reactor outlet temperature and achieve 110% design feed rates. At the increased feed rate and reactor outlet temperature, vapor load in main fractionator shed section increased, along with pressure drop.

The option to replace shed trays with structured packing or changing the tray spacing was not possible due to time limitations. Two shed trays that had the maximum vapor load were therefore removed to create more curtain area for the vapors to pass through. Of eight shed trays, two were removed, which enabled better desuperheating without entrainment.

GCU

GCU operations were limiting the feed rate at 20,000 b/sd. This led to pressure swings in the stabilizer column, inconsistent product quality of LPG and light cracked naphtha (LCN), and liquid carryover from the de-ethanizer column overhead. A detailed internal study indicated flooding of the de-ethanizer and debutanizer.

To resolve these limitations, we revamped the existing two-tower GCU configuration (de-ethanizer and debutanizer) to a three-tower configuration to handle more LPG and LCN, and improve the rvp of the cracked naphtha.

The first tower in the existing configuration (de-ethanizer) tower was converted to a primary absorber tower. The debutanizer was converted to a stripper tower. A new tower with a greater diameter and height was installed as the new debutanizer.

To overcome the problem of water accumulation in the de-ethanizer tower, a water-settling drum was provided.

This revamp, which used as many existing assets as possible, enabled us to produce about 30% more LPG and LCN from the FCC. It also helped lower the rvp of the LCN to the desired specification of 8.7 psia.

Catalyst formulation

Fresh catalyst formulation plays a vital role in determining process-unit conversions and yields, along with other parameters such as catalyst/oil ratio, reactor outlet temperature, and feed quality. The loss in unit conversion due to low catalyst circulation rate or low reactor temperature can be made up substantially by increasing the catalyst activity.

Before the revamp, fresh-catalyst activity could not increase due to severe limitations in WGC. The WGC rotor bundle revamp removed the limitation of handling LPG and gas make from the unit. We selected fresh catalyst for the revamped unit to maximize conversion and yield that match the downstream equipment recovery and capacities:

  • Fresh catalyst activity increased to 80 from 75 microactivity test.
  • Zeolite/matrix ratio increased to 85% from 70% to raise the catalyst activity.
  • Rare earth of the catalyst also increased to 1.5% from 0.5% to limit olefins in gasoline.

The new catalyst yielded significant increases in LPG and LCN yield and reduced bottoms yield.

Revamped unit performance

All the modifications were implemented during a turnaround in 2005. This resulted in increasing the yield pattern and also helped increase the feed rate to 110% of the design capacity. Table 6 shows the improvements in process unit yields.

Click here to enlarge image

LPG and LCN yields increased significantly. Bottoms yield also reduced substantially. The load to WGC increased to 28-29 tonnes/hr from 21.5 tonnes/hr. PDRCV operation improved with openings in the range of 15-25%. Reactor temperature swings have lowered to 1-2° C.

The successful Mumbai refinery revamp was coupled with a unit-specific catalyst formulation involving zeolite and matrix compositional changes, along with suitable additives.

The modifications made in stripper baffle, compressor rotor bundle, control valves, and piping increased the FCC potential by improving its throughput and yields. These low-cost solutions that cost $2 million, helped boost FCC performance by generating $0.2/bbl with a payback of 5 months.

The authors

Click here to enlarge image

Sriganesh Gandham ([email protected]) is deputy general manager, technical services, in Hindustan Petroleum Corp. Ltd., Mumbai. He has 22 years’ experience in various FCC units covering projects, operations, and technical services. Gandham holds a degree in chemical engineering from Osmania University, Hyderabad.

Click here to enlarge image

Divya Jain is deputy manager, technical services, for Hindustan Petroleum Corp. Ltd., Mumbai. She monitors FCC unit operations and oversees revamps. Jain holds a degree in chemical engineering from the Indian Institute of Technology, Kanpur, India.